Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6

Conflicts:

	drivers/net/e1000e/ich8lan.c
	drivers/net/e1000e/netdev.c

11 files changed, 163 insertions, 29 deletions

diff --git a/include/asm-mips/cevt-r4k.h b/include/asm-mips/cevt-r4k.h
new file mode 100644
index 00000000000..fa4328f9124
--- /dev/null
+++ b/include/asm-mips/cevt-r4k.h
@@ -0,0 +1,46 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2008 Kevin D. Kissell
+ */
+
+/*
+ * Definitions used for common event timer implementation
+ * for MIPS 4K-type processors and their MIPS MT variants.
+ * Avoids unsightly extern declarations in C files.
+ */
+#ifndef __ASM_CEVT_R4K_H
+#define __ASM_CEVT_R4K_H
+
+DECLARE_PER_CPU(struct clock_event_device, mips_clockevent_device);
+
+void mips_event_handler(struct clock_event_device *dev);
+int c0_compare_int_usable(void);
+void mips_set_clock_mode(enum clock_event_mode, struct clock_event_device *);
+irqreturn_t c0_compare_interrupt(int, void *);
+
+extern struct irqaction c0_compare_irqaction;
+extern int cp0_timer_irq_installed;
+
+/*
+ * Possibly handle a performance counter interrupt.
+ * Return true if the timer interrupt should not be checked
+ */
+
+static inline int handle_perf_irq(int r2)
+{
+	/*
+	 * The performance counter overflow interrupt may be shared with the
+	 * timer interrupt (cp0_perfcount_irq < 0). If it is and a
+	 * performance counter has overflowed (perf_irq() == IRQ_HANDLED)
+	 * and we can't reliably determine if a counter interrupt has also
+	 * happened (!r2) then don't check for a timer interrupt.
+	 */
+	return (cp0_perfcount_irq < 0) &&
+		perf_irq() == IRQ_HANDLED &&
+		!r2;
+}
+
+#endif /* __ASM_CEVT_R4K_H */
diff --git a/include/asm-mips/irqflags.h b/include/asm-mips/irqflags.h
index 881e8866501..701ec0ba8fa 100644
--- a/include/asm-mips/irqflags.h
+++ b/include/asm-mips/irqflags.h
@@ -38,8 +38,17 @@ __asm__(
 	"	.set	pop						\n"
 	"	.endm");
 
+extern void smtc_ipi_replay(void);
+
 static inline void raw_local_irq_enable(void)
 {
+#ifdef CONFIG_MIPS_MT_SMTC
+	/*
+	 * SMTC kernel needs to do a software replay of queued
+	 * IPIs, at the cost of call overhead on each local_irq_enable()
+	 */
+	smtc_ipi_replay();
+#endif
 	__asm__ __volatile__(
 		"raw_local_irq_enable"
 		: /* no outputs */
@@ -47,6 +56,7 @@ static inline void raw_local_irq_enable(void)
 		: "memory");
 }
 
+
 /*
  * For cli() we have to insert nops to make sure that the new value
  * has actually arrived in the status register before the end of this
@@ -185,15 +195,14 @@ __asm__(
 	"	.set	pop						\n"
 	"	.endm							\n");
 
-extern void smtc_ipi_replay(void);
 
 static inline void raw_local_irq_restore(unsigned long flags)
 {
 	unsigned long __tmp1;
 
-#ifdef CONFIG_MIPS_MT_SMTC_INSTANT_REPLAY
+#ifdef CONFIG_MIPS_MT_SMTC
 	/*
-	 * CONFIG_MIPS_MT_SMTC_INSTANT_REPLAY does prompt replay of deferred
+	 * SMTC kernel needs to do a software replay of queued
 	 * IPIs, at the cost of branch and call overhead on each
 	 * local_irq_restore()
 	 */
@@ -208,6 +217,17 @@ static inline void raw_local_irq_restore(unsigned long flags)
 		: "memory");
 }
 
+static inline void __raw_local_irq_restore(unsigned long flags)
+{
+	unsigned long __tmp1;
+
+	__asm__ __volatile__(
+		"raw_local_irq_restore\t%0"
+		: "=r" (__tmp1)
+		: "0" (flags)
+		: "memory");
+}
+
 static inline int raw_irqs_disabled_flags(unsigned long flags)
 {
 #ifdef CONFIG_MIPS_MT_SMTC
diff --git a/include/asm-mips/mipsregs.h b/include/asm-mips/mipsregs.h
index a46f8e258e6..979866000da 100644
--- a/include/asm-mips/mipsregs.h
+++ b/include/asm-mips/mipsregs.h
@@ -1462,7 +1462,7 @@ set_c0_##name(unsigned int set)					\
 {								\
 	unsigned int res;					\
 	unsigned int omt;					\
-	unsigned int flags;					\
+	unsigned long flags;					\
 								\
 	local_irq_save(flags);					\
 	omt = __dmt();						\
@@ -1480,7 +1480,7 @@ clear_c0_##name(unsigned int clear)				\
 {								\
 	unsigned int res;					\
 	unsigned int omt;					\
-	unsigned int flags;					\
+	unsigned long flags;					\
 								\
 	local_irq_save(flags);					\
 	omt = __dmt();						\
@@ -1498,7 +1498,7 @@ change_c0_##name(unsigned int change, unsigned int new)		\
 {								\
 	unsigned int res;					\
 	unsigned int omt;					\
-	unsigned int flags;					\
+	unsigned long flags;					\
 								\
 	local_irq_save(flags);					\
 								\
diff --git a/include/asm-mips/smtc.h b/include/asm-mips/smtc.h
index 3639b28f80d..ea60bf08dcb 100644
--- a/include/asm-mips/smtc.h
+++ b/include/asm-mips/smtc.h
@@ -6,6 +6,7 @@
  */
 
 #include <asm/mips_mt.h>
+#include <asm/smtc_ipi.h>
 
 /*
  * System-wide SMTC status information
@@ -38,14 +39,15 @@ struct mm_struct;
 struct task_struct;
 
 void smtc_get_new_mmu_context(struct mm_struct *mm, unsigned long cpu);
-
+void self_ipi(struct smtc_ipi *);
 void smtc_flush_tlb_asid(unsigned long asid);
-extern int mipsmt_build_cpu_map(int startslot);
-extern void mipsmt_prepare_cpus(void);
+extern int smtc_build_cpu_map(int startslot);
+extern void smtc_prepare_cpus(int cpus);
 extern void smtc_smp_finish(void);
 extern void smtc_boot_secondary(int cpu, struct task_struct *t);
 extern void smtc_cpus_done(void);
 
+
 /*
  * Sharing the TLB between multiple VPEs means that the
  * "random" index selection function is not allowed to
diff --git a/include/asm-mips/sn/mapped_kernel.h b/include/asm-mips/sn/mapped_kernel.h
index c3dd5d0d525..721496a0bb9 100644
--- a/include/asm-mips/sn/mapped_kernel.h
+++ b/include/asm-mips/sn/mapped_kernel.h
@@ -5,6 +5,8 @@
 #ifndef __ASM_SN_MAPPED_KERNEL_H
 #define __ASM_SN_MAPPED_KERNEL_H
 
+#include <linux/mmzone.h>
+
 /*
  * Note on how mapped kernels work: the text and data section is
  * compiled at cksseg segment (LOADADDR = 0xc001c000), and the
@@ -29,10 +31,8 @@
 #define MAPPED_ADDR_RO_TO_PHYS(x)	(x - REP_BASE)
 #define MAPPED_ADDR_RW_TO_PHYS(x)	(x - REP_BASE - 16777216)
 
-#define MAPPED_KERN_RO_PHYSBASE(n) \
-			(PLAT_NODE_DATA(n)->kern_vars.kv_ro_baseaddr)
-#define MAPPED_KERN_RW_PHYSBASE(n) \
-			(PLAT_NODE_DATA(n)->kern_vars.kv_rw_baseaddr)
+#define MAPPED_KERN_RO_PHYSBASE(n) (hub_data(n)->kern_vars.kv_ro_baseaddr)
+#define MAPPED_KERN_RW_PHYSBASE(n) (hub_data(n)->kern_vars.kv_rw_baseaddr)
 
 #define MAPPED_KERN_RO_TO_PHYS(x) \
 				((unsigned long)MAPPED_ADDR_RO_TO_PHYS(x) | \
diff --git a/include/asm-mips/stackframe.h b/include/asm-mips/stackframe.h
index 051e1af0bb9..4c37c4e5f72 100644
--- a/include/asm-mips/stackframe.h
+++ b/include/asm-mips/stackframe.h
@@ -297,14 +297,31 @@
 #ifdef CONFIG_MIPS_MT_SMTC
 		.set	mips32r2
 		/*
-		 * This may not really be necessary if ints are already
-		 * inhibited here.
+		 * We need to make sure the read-modify-write
+		 * of Status below isn't perturbed by an interrupt
+		 * or cross-TC access, so we need to do at least a DMT,
+		 * protected by an interrupt-inhibit. But setting IXMT
+		 * also creates a few-cycle window where an IPI could
+		 * be queued and not be detected before potentially
+		 * returning to a WAIT or user-mode loop. It must be
+		 * replayed.
+		 *
+		 * We're in the middle of a context switch, and
+		 * we can't dispatch it directly without trashing
+		 * some registers, so we'll try to detect this unlikely
+		 * case and program a software interrupt in the VPE,
+		 * as would be done for a cross-VPE IPI.  To accomodate
+		 * the handling of that case, we're doing a DVPE instead
+		 * of just a DMT here to protect against other threads.
+		 * This is a lot of cruft to cover a tiny window.
+		 * If you can find a better design, implement it!
+		 *
 		 */
 		mfc0	v0, CP0_TCSTATUS
 		ori	v0, TCSTATUS_IXMT
 		mtc0	v0, CP0_TCSTATUS
 		_ehb
-		DMT	5				# dmt a1
+		DVPE	5				# dvpe a1
 		jal	mips_ihb
 #endif /* CONFIG_MIPS_MT_SMTC */
 		mfc0	a0, CP0_STATUS
@@ -325,17 +342,50 @@
  */
 		LONG_L	v1, PT_TCSTATUS(sp)
 		_ehb
-		mfc0	v0, CP0_TCSTATUS
+		mfc0	a0, CP0_TCSTATUS
 		andi	v1, TCSTATUS_IXMT
-		/* We know that TCStatua.IXMT should be set from above */
-		xori	v0, v0, TCSTATUS_IXMT
-		or	v0, v0, v1
-		mtc0	v0, CP0_TCSTATUS
-		_ehb
-		andi	a1, a1, VPECONTROL_TE
+		bnez	v1, 0f
+
+/*
+ * We'd like to detect any IPIs queued in the tiny window
+ * above and request an software interrupt to service them
+ * when we ERET.
+ *
+ * Computing the offset into the IPIQ array of the executing
+ * TC's IPI queue in-line would be tedious.  We use part of
+ * the TCContext register to hold 16 bits of offset that we
+ * can add in-line to find the queue head.
+ */
+		mfc0	v0, CP0_TCCONTEXT
+		la	a2, IPIQ
+		srl	v0, v0, 16
+		addu	a2, a2, v0
+		LONG_L	v0, 0(a2)
+		beqz	v0, 0f
+/*
+ * If we have a queue, provoke dispatch within the VPE by setting C_SW1
+ */
+		mfc0	v0, CP0_CAUSE
+		ori	v0, v0, C_SW1
+		mtc0	v0, CP0_CAUSE
+0:
+		/*
+		 * This test should really never branch but
+		 * let's be prudent here.  Having atomized
+		 * the shared register modifications, we can
+		 * now EVPE, and must do so before interrupts
+		 * are potentially re-enabled.
+		 */
+		andi	a1, a1, MVPCONTROL_EVP
 		beqz	a1, 1f
-		emt
+		evpe
 1:
+		/* We know that TCStatua.IXMT should be set from above */
+		xori	a0, a0, TCSTATUS_IXMT
+		or	a0, a0, v1
+		mtc0	a0, CP0_TCSTATUS
+		_ehb
+
 		.set	mips0
 #endif /* CONFIG_MIPS_MT_SMTC */
 		LONG_L	v1, PT_EPC(sp)
diff --git a/include/asm-x86/uaccess_64.h b/include/asm-x86/uaccess_64.h
index 515d4dce96b..45806d60bcb 100644
--- a/include/asm-x86/uaccess_64.h
+++ b/include/asm-x86/uaccess_64.h
@@ -7,6 +7,7 @@
 #include <linux/compiler.h>
 #include <linux/errno.h>
 #include <linux/prefetch.h>
+#include <linux/lockdep.h>
 #include <asm/page.h>
 
 /*
diff --git a/include/linux/hrtimer.h b/include/linux/hrtimer.h
index 6d93dce61cb..2f245fe63bd 100644
--- a/include/linux/hrtimer.h
+++ b/include/linux/hrtimer.h
@@ -47,14 +47,22 @@ enum hrtimer_restart {
  *	HRTIMER_CB_IRQSAFE:		Callback may run in hardirq context
  *	HRTIMER_CB_IRQSAFE_NO_RESTART:	Callback may run in hardirq context and
  *					does not restart the timer
- *	HRTIMER_CB_IRQSAFE_NO_SOFTIRQ:	Callback must run in hardirq context
- *					Special mode for tick emultation
+ *	HRTIMER_CB_IRQSAFE_PERCPU:	Callback must run in hardirq context
+ *					Special mode for tick emulation and
+ *					scheduler timer. Such timers are per
+ *					cpu and not allowed to be migrated on
+ *					cpu unplug.
+ *	HRTIMER_CB_IRQSAFE_UNLOCKED:	Callback should run in hardirq context
+ *					with timer->base lock unlocked
+ *					used for timers which call wakeup to
+ *					avoid lock order problems with rq->lock
  */
 enum hrtimer_cb_mode {
 	HRTIMER_CB_SOFTIRQ,
 	HRTIMER_CB_IRQSAFE,
 	HRTIMER_CB_IRQSAFE_NO_RESTART,
-	HRTIMER_CB_IRQSAFE_NO_SOFTIRQ,
+	HRTIMER_CB_IRQSAFE_PERCPU,
+	HRTIMER_CB_IRQSAFE_UNLOCKED,
 };
 
 /*
@@ -67,9 +75,10 @@ enum hrtimer_cb_mode {
  * 0x02		callback function running
  * 0x04		callback pending (high resolution mode)
  *
- * Special case:
+ * Special cases:
  * 0x03		callback function running and enqueued
  *		(was requeued on another CPU)
+ * 0x09		timer was migrated on CPU hotunplug
  * The "callback function running and enqueued" status is only possible on
  * SMP. It happens for example when a posix timer expired and the callback
  * queued a signal. Between dropping the lock which protects the posix timer
@@ -87,6 +96,7 @@ enum hrtimer_cb_mode {
 #define HRTIMER_STATE_ENQUEUED	0x01
 #define HRTIMER_STATE_CALLBACK	0x02
 #define HRTIMER_STATE_PENDING	0x04
+#define HRTIMER_STATE_MIGRATE	0x08
 
 /**
  * struct hrtimer - the basic hrtimer structure
diff --git a/include/linux/ide.h b/include/linux/ide.h
index 1524829f73f..6514db8fd2e 100644
--- a/include/linux/ide.h
+++ b/include/linux/ide.h
@@ -366,7 +366,9 @@ enum {
 	/* Currently on a filemark */
 	IDE_AFLAG_FILEMARK		= (1 << 25),
 	/* 0 = no tape is loaded, so we don't rewind after ejecting */
-	IDE_AFLAG_MEDIUM_PRESENT	= (1 << 26)
+	IDE_AFLAG_MEDIUM_PRESENT	= (1 << 26),
+
+	IDE_AFLAG_NO_AUTOCLOSE		= (1 << 27),
 };
 
 struct ide_drive_s {
diff --git a/include/linux/ramfs.h b/include/linux/ramfs.h
index b160fb18e8d..37aaf2b3986 100644
--- a/include/linux/ramfs.h
+++ b/include/linux/ramfs.h
@@ -6,6 +6,7 @@ extern int ramfs_get_sb(struct file_system_type *fs_type,
 	 int flags, const char *dev_name, void *data, struct vfsmount *mnt);
 
 #ifndef CONFIG_MMU
+extern int ramfs_nommu_expand_for_mapping(struct inode *inode, size_t newsize);
 extern unsigned long ramfs_nommu_get_unmapped_area(struct file *file,
 						   unsigned long addr,
 						   unsigned long len,
diff --git a/include/linux/stacktrace.h b/include/linux/stacktrace.h
index 5da9794b2d7..b106fd8e0d5 100644
--- a/include/linux/stacktrace.h
+++ b/include/linux/stacktrace.h
@@ -1,6 +1,8 @@
 #ifndef __LINUX_STACKTRACE_H
 #define __LINUX_STACKTRACE_H
 
+struct task_struct;
+
 #ifdef CONFIG_STACKTRACE
 struct stack_trace {
 	unsigned int nr_entries, max_entries;